Intracellular Potassium

A Determinant of the

Sodium-Potassium Pump Rate

ANNE B. KNIGHT and LOUIS G. WELT

From the University of North Carolina School of Medicine, Chapel Hill, North Carolina27514. Dr. Knight's present address is the Department of Physiology, Duke UniversitySchool of Medicine, Durham, North Carolina 27710. Dr. Welt died 13 January 1974.At the time this article was submitted he was chairman, Department of InternalMedicine, Yale University.

ABSTRACT Normal human red cells which have had their intracellular sodium(Na,) reduced have a diminished Na-K pump rate, but only if intracellularpotassium (K0 ) is high. If most of the K, is replaced by tetramethylammoniumor choline, both ouabain-sensitive Na efflux and K influx are significantly in-creased even with Na, below normal. Cells with reduced Na, and high K,have an unchanged Na efflux if external potassium (Kt) is removed. In con-trast, low-Na, low-K cells have a large ouabain-sensitive Na efflux which showsa normal response to removal of K.et. Neither low-K nor high-K cells havean altered ouabain-sensitive K efflux. Measurement at constant low Nac andvarying K, shows the pump Na efflux to be an inverse function of K.. Thus,in low-Na cells, K, appears to act as an inhibitor of the pump. Inhibition byhigh K, can be seen even when Na, is normal. The effects attributed to K,are distinguished experimentally from other variables such as cell volume,adenosine triphosphate concentration, effects of the replacement cations, andthe method used to alter intracellular cation concentrations. A role is proposedfor K,, in cooperation with Na., in regulating the pump rate of normalhuman red cells.

INTRODUCTION

The human red blood cell maintains a high intracellular potassium (K.)and low sodium (Na,) apparently by balancing passive movements of theions with their active transport in the opposite direction. Thus the cell ex-trudes Na and accumulates K in a linked process which is sensitive to ouabainand other cardiac glycosides (Schatzmann, 1953). The component of Na-Ktransport sensitive to ouabain is termed the Na-K pump, is dependent on

adenosine triphosphate (ATP) in the cell (Hoffman, 1960), and is associated

THE JOURNAL OF GENERAL PHYSIOLOGY VOLUME 63, 1974 pages 351-373 351

THE JOURNAL OF GENERAL PHYSIOLOGY VOLUME 63 · 1974

with a ouabain-sensitive, Na-K-Mg-stimulated adenosine triphosphatase(ATPase) located in or on the cell membrane (Post et al., 1960; Skou, 1965).

Pump activity is known to be a function of its substrate ions, reachingmaximal rates at approximately 40 mM Na/liter cells (Post and Jolly, 1957)and an external K (Kext) of 10 mM (Glynn, 1956). Graphs of ouabain-sensitive Na efflux or K influx vs. Na, or K,,t give sigmoidal relationships(McConaghey and Maizels, 1962; Sachs and Welt, 1967), although thedeviation from Michaelis-Menten behavior is only apparent at very lowNa, or K,,t.

The sigmoidal behavior of the fluxes as a function of K,,t has been inter-preted both as evidence for a requirement for more than one K for inwardtransport (Sachs and Welt, 1967) and as evidence that external Na (Na..t)competes with Kext, since the curve becomes a rectangular hyperbola whenNaext is replaced by choline, tetramethylammonium (TMA), Mg, or Tris(Garrahan and Glynn, 1967 b; Sachs, 1967). The actual explanation may bea combination of the two. Precisely the same behavior has been observedwhen the Na-K-stimulated ATPase activity of fragmented membranes(ghosts) is measured as a function of K with constant high or low Na (Postet al., 1960).

In contrast to the two interpretations given to the behavior of the pumpin response to Kext, the sigmoidal response of the fluxes to varied Na, hasbeen interpreted only as evidence that more than one Na is required foroutward transport (Sachs, 1970). It has been shown, however, that theNa- K-stimulated ATPase activity of fragmented ghosts graphed as a func-tion of Na becomes a rectangular hyperbola when the K of the medium isreduced (Post et al., 1960), suggesting that the behavior of K, could resemblethat of Nat. Such an inhibitory effect of internal K on ATPase and Kinflux rates has been looked for in resealed ghosts with conflicting results,Hoffman (1962 b) concluding that high internal K affects the pump andWhittam (1962) that it does not.

Possible effects of K, on the pump have not been studied previously inintact human cells, however, and it will be demonstrated in this paper thatthe pump is inhibited by high K, particularly when the Na, is low, in amanner analagous to the inhibitory effect of Na,,t.

METHODS

Blood for all experiments was obtained in heparinized syringes from normal humanmales, and the cells were separated by centrifugation and aspiration of the super-natant plasma. The cells were always washed three times before use by resuspensionin isotonic tetramethylammonium chloride (300 mosm TMAC1) solution (170 mM),centrifugation, and aspiration of the supernatant wash.

For most experiments the cation concentrations of the cells were altered by treat-ment with p-chloromercuribenzene sulfonic acid (PCMBS), which increases the

352

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

cell's permeability to Na and K (Garrahan and Rega, 1967). Washed cells were in-cubated at a 5 % hematocrit for about 20 h at 4°C in isotonic solutions containing(mM): phosphate 20, chloride 150, adenine 3, inosine 1, glucose 14, PCMBS 0.1,varying concentrations of Na, K, and TMA to a total of 170 mM, and kanamycinsulfate (100 mg/liter). The PCMBS-treated cells were restored to approximatelynormal permeability by incubation for 30 min at 37°C in identical solutions exceptthat 2 mM cysteine replaced the PCMBS. The eysteine-treated cells were triplywashed with isotonic TMAC1 before further use. The above procedure is the methodof Garrahan and Rega (1967) with these modifications: an increased amount ofbuffer, the inclusion of energy substrates, and the use of TMA to replace some Naor K when desired.

For some experiments, the cation concentrations of the cells were altered by coldstorage for I wk at a 5 % hematocrit in similar solutions except for the absence ofPCMBS and cysteine. During this storage at 4°C the cells were resuspended in freshsolutions three or four times.

Na efflux measurements were made as follows: PCMBS-cysteine-treated or cold-stored red cells were incubated at approximately a 50 % hematocrit in solutions withthe same composition as those in which they had been pretreated, with 22Na presentin the media. The isotope-loaded cells were washed five times with isotonic TMACIand were placed in prewarmed (37C) flasks in a gyrorotary water bath at a 0.02 %or less hematocrit in media containing (mM) Mg 6, carbonate 6, phosphate 1.2,Cl 140, glycylglycine 29, glucose 11, Na 130, and K 10, with a pH of 7.4 at 370C.When it was desired to decrease either the Na or K in the medium, portions ofthese cations were replaced by osmotically equivalent amounts of TMA. Some ofthe flasks also contained 10-4 M ouabain. After a 15-20-min equilibration period,four samples were taken from each flask over a 1-h period, and after centrifugation,5 ml of the supernatant solutions were quantitatively transferred to counting tubes.A further sample was taken (at any convenient time), the cells lysed with two dropsof nonionic detergent (Acationox), and 5 ml of these hemolysates were also trans-ferred to counting tubes. Counting was done on a Packard Auto-gamma Spectrom-eter (Packard Instrument Co., Inc., Downers Grove, Ill.), usually to a minimum of10,000 counts above background, for a maximum counting error of 1 %. The left-over supernatant solutions and hemolysates were used for hemoglobin estimation bymeasuring absorption at 420 nm in a Coleman Jr. II Spectrophotometer (ColemanInstruments Div. Perkin Elmer Corp., Maywood, Ill.). Dividing the absorption ofeach supernate by that in its hemolysate gave an estimate of the fractional hemolysis(usually less than 3%). Then, kNt =- [In (I - Na*,/Na*h]/(l - fractionalhemolysis), where kN. = the rate constant for Na efflux (h-l), t = time (h), Na*, =activity of the supernatant (counts per minute per milliliter), and Na*h = activityof the whole-suspension hemolysate (counts per minute per milliliter). The rateconstants were determined by calculating the slope of the best line (by least squares) ofvalues for the right-hand side of the above equation vs. t. Then the flux, MN =(°kN.) (Na,). Since a measurement of Nac was also necessary, isotope-loaded cellswere also incubated in media identical to the flasks used for measurement of kNabut at a slightly higher hematocrit. These flasks were sampled exactly midway throughthe flux measurement, and Nac and Kc were determined on the cells, to be used inthe above equation for calculation of °MN,.

353

THE JOURNAL OF GENERAL PHYSIOLOGY · VOLUME 63 · 1974

The above procedure differs from the method of Sachs and Welt (1967) only inthe istotope-loading media used and in the inclusion of midpoint samples for Na,and K,. The former modification was necessary to maintain the abnormal cationcontent of the cells before the flux measurement, and the latter was added becausered cells tend to gain or lose ions in flux media, particularly when their Na, hasbeen raised or lowered. Therefore, a midpoint sample was considered to be morerepresentative of the actual Na, during the flux measurement than an initial Na,would be.

K efflux was measured by the same method as for Na efflux, except the cells wereloaded with 42K for 3 or 312 h.

K influx was determined by the method of Sachs and Welt (1967) with minormodifications necessitated because K influx was always measured simultaneously withNa efflux determination. Therefore, cells to be used for K influx were first incubatedfor 3 h in media exactly like the 22 Na-loading media except for the absence of 22Na.Flasks containing solutions as described for Na efflux, but containing 42K, wereprewarmed in a gyrorotary water bath, and washed cells were added to the flasksto approximately a 5 % hematocrit. Two samples were taken I h apart; the super-natant solutions were saved and diluted appropriately for counting; the cells werewashed rapidly three times with cold isotonic TMACI solution and lysed with twodrops of nonionic detergent. The lysed cells were diluted to a convenient volume(usually 6 ml) and exactly 5 ml of each hemolysate were transferred to countingtubes, the remainder being used for hemoglobin determination by the cyanomethe-moglobin method. The diluted supernatant solutions and the hemolysates werecounted as for the Na efflux method, with correction made for counts lost by decaymade when necessary. Then, iMK = (AK*¢)/(K*8/K.), where iMK = the K influx(millimoles per liter cells per hour), AK*, = the increase in activity (counts perminute per milliliter) in the cells over exactly 1 h, and K*,/K, = the specific ac-tivity of the supernatant solution. For the calculation it was necessary to know thevolume of cells hemolyzed and counted; therefore, hemoglobin (by the cyanomethe-moglobin method) and hematocrit were determined on a sample of cell suspensionused for the flux, and v = (volh · hgbh)/(hgbrs/hct,r), where v = ml of cells counted,volh = ml of hemolysate counted, hgbh = g/ml of hemoglobin in the hemolysate,hgb,. = g/ml hemoglobin of the reference suspension of cells, and hct,, = fractionalhematocrit of the reference suspension.

For all flux measurements, the pump was taken to be the flux in the absence ofouabain minus the flux in the presence of 10-4M ouabain. Na and K were measuredsimultaneously on an Instrumentation Laboratories flame photomometer (Instru-mentation Laboratory, Inc., Lexington, Mass.) with Li as the internal reference.For determination of Na, and K,, the cells were always washed three times withNa- and K-free isotonic TMAC1 solution, and Na, K, (in duplicate), and hemato-crit (in triplicate) were measured on the washed-cell suspension. Then, millimolesper liter cells of Na (or K) = mM Na (or K) in the suspension - hematocrit ofthe suspension.

All results presented are the means of at least duplicate flux measurements whichdiffered from each other by no more than 5 %. The single exception is Fig. 5 inwhich single data points from three entirely separate experiments are given.

Intracellular ATP concentration was measured on a cell hemolysate by the method

354

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

of Dufresne and Gitelman (1970), a semiautomated procedure in which ATP isreacted with a luciferin-luciferase firefly extract, and the resulting light flashes arecounted in a Packard Tri-Carb liquid scintillation spectrometer.

Percent cell water was usually determined by weighing red cell suspensions ofknown hematocrit both wet and after being dried for 24 h at 100°C. Then, % cellwater = 100 [wet-dry)/(wet-tare) 100-(100 - % hct)]/% hct.

Cell chloride concentration was measured on cells packed at about X30,000 gfor 5 min in specially constructed lucite tubes with a -cm diameter well (the re-maining packed cells being used for water determination). The cells were hemolyzedand diluted 20-fold with water: 1 ml of this hemolysate plus 3 ml water was treatedwith 0.5 ml of 0.7 N ZnSO4 - 0.25 N H2 SO4 and 0.5 ml of 0.75 N NaOHto deproteinize the hemolysate, and I ml of the resulting clear supernatant solutionwas reacted with 3 ml chloride reagent (0.1 N HNOs - 10% acetic acid) and ti-trated on a Buchler Cotlove chloridometer automatic titrator (Buchler InstrumentsDiv., Nuclear-Chicago Corp., Fort Lee, N. J.) with appropriate standards. Chloridewas also determined on the solution in which the cells had been packed (0.1 mlsupernate plus 4 ml chloride reagent). The chloride ratio (Cl, per kilogram cellwater/Clext per kilogram external water) was then used to calculate the membranepotential using the Nernst equation, Em = -- 0.058 log (Cl,/Clex), where Em is thetransmembrane potential in volts.

Oxmolality was measured by freezing-point depression using an Advancedosmometer (Advanced Instruments, Inc., Needham Heights, Mass.). 22Na and a2Kwere obtained as the chlorides in water from International Chemical & Nuclear Cor-poration, Burbank, Calif. or from Cambridge Nuclear Radiopharmaceutical Corpora-tion, Cambridge, Mass. Choline chloride (Nutritional Biochemicals Corporation,Cleveland, Ohio) was purified before use by recrystallization from 5% ethanol.Sigma Chemical Co., St. Louis, Mo. was the source of adenine, inosine, PCMBS,L-cysteine, and ouabain. Kanamycin sulfate solution was from Bristol Laboratories,Syracuse, N. Y.

RESULTS

Effect of Intracellular K on Na Efflux and K Influx in Low-Na Cells

Normal cells were divided into two groups, and both were treated withPCMBS solutions containing either 170 mM K or 30 mM K plus 140 mMTMA. Both groups of cells thus became very low in Na, but with intracellularK either slightly above normal or about 60% below normal. Ouabain-sensitive Na efflux was then measured on five such paired groups of cells,and K influx was measured simultaneously in three of the experiments. Theresults are shown in Fig. 1. Both ouabain-sensitive Na efflux and K influxare significantly reduced in low-Na, high-K cells. This reduction in bothhalves of the coupled pump cannot be explained by the lowered Na, alonebecause low-Na cells with the K. also reduced show a very large pump flux.Thus both aspects of the coupled pump are reduced when only the Na islowered, and increased when the K, is also lowered. The ouabain-sensitiveNa efflux from low-Na cells shows an inverse response as K. is varied (Fig. 2)both in the presence of saturating Kext (10 mM) and in the absence of Keit.

355

Nc' 3-4 mmol/liter cellsKc · 110-1

30mmol/liter cells

Kc 30-40mmol/liter cellslIn .. X _ ,,Rue -IoV mMKext IOmM

V

71

V/

No=3-4 mmol/liter cells017

0

CCC

EE

.c_Iv

LHi

(I-Z

zW

oL

Z

0

FIGURE 1. Effec of intracellular K (K,) on ouabain-sensitive Na efflux (I a) and Kinflux (1 b). Cells were treated with PCMBS, followed by cysteine, to alter intracellularcation concentrations, as described under Methods. TMA was used to replace K in thelow-K cells. In this and the following figures, the standard flux medium was used,with only the Na, K, and TMA of the media being varied. This standard mediumcontained (millimoles per liter) 6 MgCOa, 1.2 phosphate, 29 glycyl-glycine (pH 7.4),140 C1, II, glucose, and Na, K, and TMA as described in each figure.

356

F

_Z.1

1.1

1.6

0C-

EE§

I-

Z

U)

z

w

I?z

0m

z

0

1.2

1.0

0.8

0.6

0.4

VgA

0.2

0o . V .A VZ il -.

~A

I I

gIns

I.

= 1.6

1.42 1.4

E

Z IO

> 0.8

I

S 0.6z

o 0

0

0.2

,1.4

.0

-

l 1.2

z

w0

0

Z

> .8

<n

n 0.6

z

m<0.4o

0.2

O

®cells

20 40 60 80 100 120Kc mmol/liter cells

140

Noa = 3.0-3.9 mmol/liter cellsNa.et = 130 mMKe t = 0 mM

20 40 60 80 100Kc mmol/liter cells

120 140

FIGuRE 2. Response of ouabain-sensitive Na efflux to K at constant low Nac, withsaturating Kxt (2 a) or zero Kt (2 b). Cells were PCMBS-treated to alter intracellularcation concentrations, TMA being used to replace K.

357

I I I I I I . . I

I I I I I I I

O.8

16

THE JOURNAL OF GENERAL PHYSIOLOGY VOLUME 63 · 1974

Response of Low-Na Cells to External K

The ouabain-sensitive Na efflux from normal red cells, as reported by Gar-rahan and Glynn (1967 a) is reduced by about one-third when Ket is re-duced from saturating concentrations to zero as long as Naext is kept high.With zero Kext, the pump is believed to perform Na-Na exchange insteadof Na-K exchange, and the unidirectional efflux of Na is one-third slowerwhen the cells are exchanging Na than when K is available externally.

Comparison of Fig. 2 a and 2 b suggested that the Kc has some effect onthis sensitivity of the cells to Kext. Data relevant to this point are shown inTable I. Both groups of cells in this table have a low Na0, but those witha high Kc show the same low ouabain-sensitive Na efflux whether or not Ke,,tis present. In contrast, low-Na cells with a low Kc show a sensitivity to Kexttypical of normal cells.

Table I also includes cells in which K. was replaced by choline, with ex-actly the same results as when TMA was used.

The very small ouabain-sensitive Na efflux of low-Na, high-K cells, com-

EFFECT OF

TABLE I

REMOVAL OF Kxt ON oMN-s

oMOf-

Nac Ko Kext = 10 mM Ket= 0 mM Difference

(mmol/liter cells)

I High Kc cells2.3 117 0.11 0.19 -0.082.9 129 0.10 0.14 -0.042.9 127 0.26 0.27 -0.013.4 129 0.43 0.40 0.033.4 120 0.27 0.33 -0.063.7 117 0.33 0.39 -0.064.2 125 0.46 0.44 0.02

II Low K cells3.2 36 1.38 1.22 0.163.4 37 1.50 1.12 0.383.5 38 1.73 1.14 0.594.2 50 1.74 1.08 0.66

III2.8 23 1.61 0.71 0.903.4 28 1.72 1.06 0.66

Cells were treated in PCMBS solutions, as described under Methods, witheither high K (group I), high TMA (group II), or high choline (group III),to obtain cells with low Nac and high or low Kc. Na-efflux measurements(millimoles per liter cells per hour) were carried out in the standard fluxmedia, with Na.xt = 130 mM and either Kext or TMAext = 10 mM.

358

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

bined with the lack of response to Kt of such cells, suggests the possibilitythat the active Na-K transport of the cells is almost entirely turned off bythe high K, and the small transport ability remaining is equally as capableof performing Na-Na exchange as Na-K exchange.

Studies of Net Na Movement

Low-Na cells are not in a steady state, and since the determination of Naefflux rate by tracer measurements is based on the assumption that the mid-point Nac represents the average Nac during the flux measurement period,a number of nonisotopic measurements of Na movement were made on low-Na cells, both to check the validity of the results already presented, and toconfirm that a midpoint Nac measurement represents a true mean in suchcells.

Fig. 3 a shows such a study on cells with an initially low Na, and highKc, where the total Na. was measured in samples taken over a 2-h period. Inthese cells neither the removal of K,,.t nor the presence of ouabain signifi-cantly affected the initial rate of accumulation of Na, which agrees with the22Na measurements. As the Na, increased, the net Na accumulation rateshowed a slight ouabain-sensitive decrease in the presence of Kt, indicatingthat the pump is accomplishing a small net Na extrusion as the Na, approachesnormal. Furthermore, the gain of Na is linear under all conditions for thefull 2-h period, making a midpoint measurement of Na a valid estimate ofthe mean concentration during the flux period. As previously noted, isotopicfluxes were measured over 1 h, starting 15 min after addition of the cells tothe media.

Fig. 3 b shows similar data for cells with an originally low Nac and lowK., and it can be seen that both ouabain and Ket have significant effectson these cells. For uninhibited cells, the removal of K,,t increases the rateof Na accumulation, and the addition of ouabain increases it further. Inthe absence of Ket the cells show no effect by ouabain, as expected, sinceNa-Na exchange does not result in net Na extrusion. Thus, these results agreewith the 22Na measurements on low-K cells. In addition, the gain of Na islinear for the full 2 h, except in the case of uninhibited cells in the presenceof Kext, and in this case linearity is approximated over the period of concernfor tracer measurements.

Effect of K, on K Efflux

Although the K efflux observed from normal human red cells is movementalong its electrochemical gradient, a small portion of this efflux is inhibitedby ouabain and has been attributed to K-K exchange through the pump(Glynn and Luthi, 1968). This ouabain-sensitive K efflux can also be ob-served in the absence of K,.t in a high-Na medium, and under these condi-

359

f. -a Kix t 10mIM WITH OUABAIN- Kext 10 mM NO OUABAIN

&r--A Kext O. WITH OUABAIN, o--o Kext O. NO OUABAIN

INITIAL Kc ' 120-122 mmol/liter cells

/,,

E

30 60 120MINUTES

'u

9

~a

' N

EE

I-

z0

Z50

M4

0),

-Jwu

2

II

n

-- 0 Ke t = IOmM WITH OUABAIN

*-* Kext 1OmM NO OUABAIN

'- "~ K.ex =

O. WITH OUABAIN

O---0 KeXt =

0. NO OUABAIN

INITIAL Kc = 25-28 mmol/liter cells

Ne.xtt 130 mM /

/

/,/) / '.

30 60 120MINUTES

FIGURE 3. Comparison of net Na gain by low-Na cells, with either high (3 a) or low

(3 b) K. PCMBS-treated cells. Each graph is representative of three experiments.

360

Q

7

O 6EE

Z 4

u5

zo

0

3

I.-

_1J

<.

{'

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

tions it is thought to represent an entire reversal of the pump (Glynn andLew, 1970). In both cases, the K efflux has been demonstrated to be a spe-cific one and not simply a lack of discrimination between Na and K by thecarrier (Glynn et al., 1970). This observation makes it unlikely that the in-hibitory effects of K, observed here are due to a simple substitution and trans-port by K at Na sites on the carrier. In line with this idea, Fig. 4 shows that theouabain-sensitive portion of the K efflux from low-Na cells is not decreasedwhen K, is lowered to 24 mM cells.

It may be that a much larger reduction of K, would affect the ouabain-sensitive K efflux from low-Na cells, but it seems most likely that a high K,simply interferes with the outward transport of Na from such cells withoutitself being carried at a faster rate.

Effect of K, at Normal and High Levels of Na,

The question next arises whether the high K, of human red cells is inhibitoryto the pump at Na, levels in the normal range. All prior investigations of theeffect of varying Na, on the pump rate in human cells have been done bysimultaneously varying the K, in a reciprocal manner. Thus, in previouswork, the K, has been low only when the Na, has been high.

Fig. 5 shows experiments measuring ouabain-sensitive Na efflux as a func-tion of Na,, with the K, held constant at two levels. The Na.1 t was largelyreplaced by TMA in order that the Na, might remain low throughout theperiod of measurement. The K, of cells represented by the solid line is thesame as in normal cells and the Na,'s include the entire normal range. Fig.

0

0

E

-f

zI

X

Kext 10 mM Kext'O Kext - IOmM Kext=O

Kc=118-124 mmol/liter cells Kc= 24 - 27mmol/liter cells

FIGURE 4. K efflux from low-Na cells with high (left group) or low (right group)K,. Naet = 130 mM. Each bar represents total K efflux, with subdivisions for ouabain-sensitive and ouabain-insensitive components. PCMBS-treated cells.

361

THE JOURNAL OF GENERAL PHYSIOLOGY VOLUME 63 · 1974

5 a thus indicates that the normally high K, of human red cells is inhibitory tothe pump when compared to cells with a normal Nac and a reduced K,.In Fig. 5 b, the rate constant for ouabain-sensitive Na efflux is plotted againstNa, and it can be seen that the peak of the curve is shifted to the left inlow-K cells, as one expects when the concentration of an inhibitor is reduced.

-Kc 104-106'"-- Kc 15-30

Next = 6

Kext = 10

x

0

15 16 17

Noc mmol/liter cells

0- Kc 10

S " ---- Kc 1500_. ^ xftft' Naext

o Kext

0xl ~ ~ ~ ~ xAgo O 4

II 0

| - ~~o o~

4-1065-30

=6t10

x

x III/ I I

I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Na c mmol/liter cells

00

Zz

E

O

z

,)z

C,

0

u.(

0.6

0.5

0.4o kO-s

kNa

0.3

0.2

0.1

00

" ......... I I I I I I

362

(

s

Z

(

n . __

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

0oM-SKN

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Nac

FIGuRE 5. Ouabain-sensitive Na efflux from PCMBS-treated cells with varying Na,

and two levels of K, (millimoles per liter cells). Results (of three experiments) ex-pressed as efflux (5 a), as the rate constant (h-1) for efflux (5 b), and as a double-recip-rocal plot (5 c).

Fig. 5 c is a double reciprocal plot of the data of Fig. 5 a: /flux versusI/Na, at two K, levels. The line for high-K cells has a much larger slopeand curvature than that for low-K cells.

Several measurements were made of the ouabain-sensitive Na efflux withNa, between 49 and 97 mmol/liter cells, well above the level needed formaximal velocity of the pump (Post and Jolly, 1957). At the same time, theK, was varied from 4 to 55 mmol/liter cells. The results (Table II) showno significant effect on Vm, by K, for any K, greater than 4 mmol/litercells. When the K, is four or less, however there is an apparent reduction inVm,. In no case is Vm,, increased by lowering K, in the range studied.

363

_ _

THE JOURNAL OF GENERAL PHYSIOLOGY VOLUME 63 · 974

TABLE II

EFFECT OF Kc ON Vma,

Nac Kc °MN-as

(mmol/li tr cells) (mmol/liter cells/h)

I 56 55 4.3953 31 5.0949 20 5.0257 4 4.35

II 56 52 4.6350 27 4.3445 17 4.3352 3 3.26

III 64 47 5.2484 28 5.5297 4 3.9678 4 4.72

IV 59 48 6.1183 27 6.5695 4 4.4176 4 4.78

Ouabain-sensitive Na efflux from PCMBS-treated cells with Na, above 45mM cells. Naext = 130 mM, Kaxt = 10 mM. Groups Ic and III were fromthe same donor. Mean of all cells with Kc 30 or above = 5.08 4 0.3 SEM.Mean of cells with K, less than 10 = 4.26 4 0.22 SEM. P = 0.025 that cellswith K, < 4 are different from cells with Kc, 17.

Effects of Other Variables

It was considered possible that any of several other variables might actuallybe responsible for the inhibitory effects which we have been attributing tothe K,. Possible effects of the replacement cation and the nonsteady-statesituation have been mentioned previously. Remaining possibilities include(a) effects of PCMBS or cysteine, (b) cell volume or membrane potentialchanges, and (c) different levels of ATP in high- and low-K cells.

The first-listed variable is excluded by the data shown in Fig. 6. Thisfigure differs from Fig. 5 in that the cation concentrations in the cells werealtered by cold storage instead of by PCMBS treatment: The ouabain-sensitive Na efflux is plotted against Na0 at two K. levels in Fig. 6 a and as adouble reciprocal plot in 6 b. Cold storage is not as effective as PCMBStreatment in reducing the K., but the effect of the K, on the Na efflux is thesame in either case. Thus the inhibitory effect of a high K, cannot be at-tributed to effects of PCMBS or cysteine. It should also be noted that thebehavior of the flux as a function of Na, is unaffected by the presence of high(130 mM) or low (6 mM) Naet, given a constant K 1,t of 10 mM (cf. Figs. 5and 6), although the absolute magnitude of the flux may be affected.

3°4

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate 365

The second variable which might mimic an inhibitory effect of high K,,that a coincident volume change in low-K cells might increase the pumprate, is considered in Fig. 7. Cell shrinkage, of necessity, would increase theeffective Na,, tending to stimulate the pump. It was, in fact, observed thatfor cells with the same Na, (per liter cells), the cell water was higher in high-K cells and was reduced in low-K cells (this is shown in Table III). However,

= K 10I0o K 55

0 1 2 3 4 5 6 7

Nacmmol/liter cells

8 9 10 II

K 108-110Kc 55-69

®

Nac

FIGURE 6. Ouabain-sensitive Na efflux from cold-stored cells with varying Nac andtwo levels of Kc (millimoles per liter cells). Results expressed as efflux (6 a) and as adouble-reciprocal plot (6 b).

EE

Iz'C

-

o)z

z

m

o

3.U

2.5

2.0

1.5

1.0

0.5

C

M O-SMN

1.0

--

I II I I III

0

THE JOURNAL OF GENERAL PHYSIOLOGY ·VOLUME 63 ·1974

when the ouabain-sensitive Na efflux and the Nac are expressed in terms ofthe cell water (Fig. 7), it can be seen that the inhibitory effect of a high K.on the pump is still apparent.

Paralleling the observed volume changes, it was also found that therewas a difference in intracellular chloride between low- and high-K cells.Table III shows the chloride ratios of normal cells and of low-Na cells withhigh and low K, the membrane potentials calculated from the C1 ratios,and the % cell water. It can be seen that the C ratio is reduced in low-Kcells and the membrane potential is correspondingly more negative.

The final potential source of an apparent effect of K. on the pump is thecells' ATP concentration: If high- and low-K cells had significantly differentATP levels it could account for the observed differences in their ouabain-sensitive Na effiux rates. Table IV shows ATP measurements on high- andlow-K cells with various Na,'s, including Nac above that needed for Vm.x.It is evident that the energy substrates provided are sufficient for the cellsto maintain ATP within the normal range irrespective of their Na and K

-

oa

EE

0o

0

25 30

Nac mmol/kg cell water

FIGURE 7. Replot of results shown in Fig. 5, with the flux and Na, expressed in termsof intracellular water.

366

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

TABLE III

CHLORIDE RATIO, CELL WATER, AND MEMBRANE POTENTIAL OF NORMALAND LOW-Na CELLS

Cl¢/Clezt Em H20

(mV) (%)

Untreated cellsNa, = 6-7.8

K, = 98-104 0.6340.01 -11.840.06 63.2-0.4

Low-Na, high-K cellsNa, = 0.64.0

K¢= 114-123 0.6440.04 -11.24-1.0 68.140.2

Low-Na, low-K cellsNac = 0.6-3.6

Kc = 20-40 0.4340.03 -21.24-1.3 57.0-0.5

Chloride ratios (millimoles Cl per kilogram cell water . millimoles Cl per kilogram extra-cellular water), calculated membrane potentials, and percent cell water for normal, high-,and low-K cells. Each number represents mean - SEM of eight determinations. Cl and H2 0measured on PCMBS-treated cells packed in 170 mM TMA Cl. Cl ratio and Em not signifi-cantly different for untreated cells compared with high-K cells (P > 0.5). C1 ratio and Emsignificantly different for low-K cells compared with either normal or high-K cells (P < 0.001).Cell water values for all three groups are significantly different from each other (P < 0.001).Na, and K, expressed as mmol/liter cells.

concentrations. Therefore, a difference in ATP cannot be responsible forthe observed differences in pump activity.

DISCUSSION

The behavior of the Na-K pump of intact human red cells in response tochanges of K, can best be explained by interpreting the K, as a competitiveinhibitor of the pump. Such an interpretation fits in well with the pre-viously demonstrated inhibition of the pump by Naext (Sachs, 1967). Con-sideration of K, as a similar (but not identical) inhibitor is further supportedby the response of the Na- K-stimulated ATPase of ghosts to variations ofNa and K, which have been observed by previous investigators.

The results of studies with the Na, and K, varied independently (Figs.5 and 6) distinguish the known stimulatory effect of increasing Na, from theinhibitory effect of high K,: either an increase of Na, or a decrease of K,raises the pump rate. Conversely, while a decreased Na, reduces the pumprate, a simultaneous reduction of K, raises the pump rate significantly,whether Na efflux or K influx is being measured (Fig. 1). In fact, if the Na, ismade low and held constant while the cell K is varied, the flux rate is aninverse function of the Kc (Fig. 2). If the Na, is held low, only when the K,is reduced such that the ratio of K, to Na, is approximately the same as

367

THE JOURNAL OF GENERAL PHYSIOLOGY ·VOLUME 63 ·1974

TABLE IV

ATP CONTENT OF HIGH- AND LOW-K CELLS

Group Nac Kc ATP Group Nac Kc ATP

(mmol/likr cells) (mmol/liter cells)

I 1.3 119 1.2 IV 56 55 1.50.9 46 1.7 53 31 1.7

49 20 1.7II 2.8 127 1.6 57 4 1.5

3.0 27 1.8 56 52 1.42.8 128 1.6 50 27 1.23.0 24 1.7 45 17 1.5

52 3 1.3III 1.0 100 2.0 64 47 1.6

5.5 104 2.3 84 28 1.711.0 104 2.3 97 4 1.60.9 29 2.4 78 4 1.55.0 37 2.1

12.0 37 2.3 V 3.5 118 1.23.5 118 1.34.5 57 1.54.2 56 1.6

ATP concentrations in cells with varied Nae and Kc . Flux measurements for cells in group IIare included in Fig. 4, for group III in Fig. 5, and for group IV in Table II. Cation contentin all groups was altered by PCMBS treatment, except group V, which was cold stored. Meanof high-K cells = 1.69 4 0.16 SEM (groups I, II, III, IV). Mean of low-K cells = 1.89 4 0.12SEM. P > 0.15.

cells with a normal K, and Nac (about 10 to 1) does the pump rate increaseto approximately the rate seen in normal cells.

A combination of a very low Na. and a high K, appears to shut off thepump to such an extent that the normal response of the Na efflux to stimula-tion by Ket disappears (Table I). Nor do such cells have an increased Kefflux (Fig. 4). Since ouabain-sensitive K efflux represents reversal of at leastpart of the normal pump operation (Glynn and Lew, 1970), this can be in-terpreted as another indication that the pump itself is very nearly inacti-vated in all of its operations. The cells in these experiments probably con-tained adequate orthophosphate to allow K efflux to occur, since they hadbeen exposed to 20 mM phosphate and only 1 mM glucose. Such pretreat-ment maintained ATP levels (Table IV) and should have allowed intra-cellular phosphate to remain at least normal, although it was not measured.This is of interest because Glynn et al. (1970) showed that orthophosphateand high Na,,t or K,,t were the major requirements for K efflux to occur.These minimal requirements were apparently met here since there was anobservable K efflux, but it is possible that to show an increased K efflux itwould be necessary to actually elevate the internal phosphate. In any case,the ouabain-sensitive K efflux is not affected by K, in the same range and in

368

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

similar conditions to those in which it obviously affects the Na efflux andK influx of low-Na cells.

The following were shown not to be responsible for the evident inhibitoryeffect of high Kc on the pump: (a) volume changes that occur when the Kcis changed (Fig. 7); (b) the intracellular ATP concentration (Table IV);(c) nonsteady-state situation (Fig. 3); (d) effects of PCMBS or cysteine (Fig.6); or (e) the choice of a replacement cation. This last point is based notonly on the use of two replacement cations in these experiments (Table I),but also on the observations of other investigators studying the effects ofvarious cations on the pump and ATPase. The response of the pump toKet is of the same form in media containing high TMA, Mg, choline, orMg-sucrose (Sachs, 1967; Sachs, 1970; Garrahan and Glynn, 1967 a; Priest-land and Whittam, 1968). Similarly, the ATPase of resealed ghosts hassimilar properties when the external medium contains Mg, Tris, choline,or TMA (Hoffman, 1962 a; Whittam and Ager, 1964; Schatzmann, 1965;Whittan, 1962). In the case of the ATPase of fragmented ghosts, it has beenfound that the response of the ATPase to variations of Na and K is no different ifTMA or choline is used to bring the medium to isotonicity or if the mediumis left hypotonic (Smith and Welt, 1971; Robinson, 1970). Furthermore,there have been several direct demonstrations of an inhibitory effect ofhigh K on the ATPase (as well as a complementary inhibition by high Na)on fragmented red cell ghost ATPase (Post et al., 1960; Dunham and Glynn,1961) and no nerve and kidney microsomes (Skou, 1957, 1960; Green andTaylor, 1964; Robinson, 1970; Ahmed et al., 1966); in all of these experi-ments, apparently no replacement cation was used as the Na and K of themedia were varied.

Although volume changes leading to changes in the effective concentra-tion of Na or K in the cells have been excluded per se as an explanation forthe effects of Kc (Fig. 7), it was observed that low-K cells were not onlyshrunken but also had a reduced chloride ratio, and therefore a more nega-tive calculated membrane potential (Table III). A possible explanationfor both the reduced C1 in the cells and the reduced volume is that, whilePCMBS increases the cells' permeability to Na and K (Garrahan and Rega,1967) so that they lose K in a low-K medium, they apparently do not take

up enough TMA to replace all of the lost K. The cells thus lose water, pack-ing the hemoglobin more closely. Gary-Bobo and Solomon, (1968) haveobserved that the net charge on the hemoglobin molecule becomes morenegative as its concentration increases; thus shrunken cells tend to lose C1,decreasing the C1 ratio. The resulting more negative membrane potentialcannot account for the increased pump rate in low-K cells since the Naefflux portion of the reaction goes against the electrical gradient. However,the reduced C1 ratio also indicates an intracellular pH reduction, and the

369

THE JOURNAL OF GENERAL PHYSIOLOGY VOLUME 63 1974

resulting pH gradient cannot be discounted as being the actual mechanismfor the increased pump rate of low-K cells.

As referred to above, there have been several direct demonstrations ofinhibition of the ATPase by high Na or K. In addition, several investigatorshave postulated that the high Kc of intact cells would be found to be in-hibitory to the pump, at least in low-Na cells, either on the basis of indirectevidence (Maizels, 1968; Skou, 1965) or on the basis of models of the pump-ATPase system (Post et al., 1960; Opit and Charnock, 1965), or on both(Hoffman and Tosteson, 1971). An attempt to demonstrate an effect of in-ternal K on the ATPase of resealed ghosts was made by Whittam (1962).However, since resealed ghosts undergo a pronounced volume change duringtheir preparation, it is difficult to compare the internal ion concentrationsin his experiments to concentrations in intact cells. The internal sodium con-centrations in the ghosts were rather high, though still in the range wherethe internal Na did affect the activity, and no influence of internal K wasseen over a fivefold range. He concluded that the ATPase activity was afunction only of the internal Na and the external K.

That same year, however, Hoffman (1962 b) measured Na effiux from re-sealed ghosts which had been hemolyzed either in water or in hypotonicKCL, and observed a higher strophanthidin-sensitive Na efflux from ghostsprepared in the former manner. Although the experiment was complicatedby tonicity effects and by heterogeneity of the ghost preparations, it stronglysuggests an inhibitory effect by internal K.

The fact that the inhibitory effect of K, on the pump is modified by asufficiently high Nac is the probable explanation for Whittam's failure toobserve inhibition by internal K in ghost preparations. In fact, the Na, mustbe normal or slightly below normal before the inhibitory effect of a high Kcbecomes readily observable (Fig. 5). At the other end of the scale, with Na,well above saturating levels of about 40 mmol/liter cells, reducing the K, toofar also lowers the flux rate (Table II). This apparently anomalous effect ofK, recalls the apparent reduction of the flux in media containing saturatingKext which occurs when all or most of the Naxt is replaced by other ions(Hoffman and Kregenow, 1966; Lubowitz and Whittam, 1969). Sachs(1970) has suggested that most of the pump increase due to Na,,, can be ex-plained by an increasing Nac during the flux measurement in the presenceof high Naext. Because Kext is already saturating, a similar explanation can-not be applied to the apparent Vma alteration by K,. It is, however, possiblethat a minimal amount of K, and Na,,t are necessary for the pump to retainits normal responsiveness to Ket and Na,. It appears that both the ratios ofK, to Nac and of Naext to Kot, as well as their absolute amounts, determinethe pump rate.

At a primary level, the data presented here suggest that the inhibition of

370

KNIGHT AND WELT Intracellular K, a Determinalt of Na-K Pump Rate

the pump by high K, is of the competitive type (affecting the slope of adouble-reciprocal plot), and, in fact, that it resembles the known competi-tion by Naext (Sachs, 1970). It cannot be excluded, however, that the in-hibition is of the mixed type (affecting slope and intercept); distinguishingbetween these alternatives by direct means may actually not be possible inintact cells, because it would be necessary to increase the Nac above 40mmol/liter cells, keeping the Kc high at the same time. It should be notedthat reduction of K, to 55 mmol/liter cells is enough to stimulate the pump(Fig. 6). Since a K, of 55 was the highest obtained with Na0 over 40 (Ta-ble II), the data of Table II are not sufficient to determine conclusivelywhether or not K. affects Vmax with respect to Na,.

Unfortunately, the mechanism of the pump is as yet too uncertain toenable one to make a decision between Kc as an allosteric effector and com-petitive inhibition by K, at more than one substrate site. It seems apparentthat there are at least two sites for K,,t (Sachs and Welt, 1967). Similarly,evidence has been presented by Sachs (1970) for at least two sites for Na,;the form of Fig. 5 b supports this observation, since it would be difficult toaccount for the biphasic form of the curve of rate constant vs. Na, withoutat least two Na. sites. Robinson (1970) also supports a two-Na site inter-pretation from work on the ATPase. K, apparently interacts either directlyat the two Nac sites, or in an allosteric manner at separate sites. Assumingthat there are two sites on the carrier for the active effiux of sodium, onecould suggest two possibilities for the inhibitory action of K,. Both viewsimply a competition between Na and K for the Na sites and that with higherK, more Na sites are preempted by K. The first view is that having pre-empted a site the K is then ejected from the cell. This appears to be un-supported by the data since the ouabain-sensitive K-K exchange and theouabain-sensitive K effiux into a K, medium is not different in high- andlow-K, cells. The second view is that when K preempts Na sites the carrieris deterred in such a fashion as to diminish the rate of pump activity withoutinfluencing K efflux. This view is still compatible with these data. It is,in any case, evident that K. is an important determinant of the pump rate.

We propose that the inhibitory effect of a high Kc demonstrated here ispartially responsible for controlling the physiological level of Na and K innormal red cells. An increased Na, serves as its own regulator by directlystimulating the pump rate, reducing the Na0 . Since the normal high K,of human red cells is in the inhibitory range, and since it exerts its inhibitoryeffect most strongly when the Na0 is low or normal (Figs. 5 and 6), the K,would be important in preventing the Nac from dropping below normal.If the Na0 were to decrease slightly, the pump rate should be reduced notonly by the reduction of substrate concentration but also by inhibitoryaction of the K,, allowing a gain of Na by diffusion. The cell would thus be

37I

THE JOURNAL OF GENERAL PHYSIOLOGY · VOLUME 63 · 1974

more sensitive to slight reductions of Na. than if the K, were to have no effecton the pump. A similar argument would apply to the controls exerted by theratio and concentrations of the external ions.

Dr. Knight's work was supported in part by USPHS Training Grant AM05054 and USPHS Re-search Grant HE01301.Parts of the present work have appeared in preliminary form: Knight, A-, R. C. Taylor, and L. G.Welt. 1971. Fed. Proc. 30:314 and Knight, A. and L. G. Welt. 1972. Int. Congr. Neph. Abstr. MexicoCity. 5:138.This represents work done by Dr. Knight in fulfillment of the requirements for the Ph.D. in Bio-chemistry at the University of North Carolina School of Medicine May, 1972.

Note Added in Proof Since the submission of this manuscript, a paper has appeared

(R. P. Garay and P. J. Garrahan. 1973. The interaction of sodium and potassiumwith the sodium pump in red cells. J. Physiol. (Lond.). 231:297) which also suggests

that high intracellular K competes for pump sites with internal Na in the normalrange.

Received for publication 1 March 1973.

REFERENCES

1. AHMED, K., J. D. JUDAH, and P. G. SCHOLEFIELD, 1966. Interaction of sodium and potas-sium with a cation-dependent adenosine triphosphatase system from rat brain. Biochim.Biophys. Acta. 120:351.

2. DUFRESNE, L., and H. J. GITELMAN. 1970. Semiautomated procedure for determinationof adenosine triphosphate. Anal. Biochem. 37:402.

3. DNHAM, E. T., and I. M. GLYNN. 1961. Adenosine triphosphatase activity and the activemovements of alkali metal ions. J. Physiol. (Lond.). 156:274.

4. GARRAHAN, P. J., and I. M. GLYNN. 1967 a. The behavior of the sodium pump in redcells in the absence of external potassium. J. Physiol. (Lond.). 192:159.

5. GARRAHAN, P. J., and I. M. GLYNN. 1967 b. The sensitivity of the sodium pump to ex-ternal sodium. J. Physiol. (Lond.). 192:175.

6. GARRAHAN, P. J., and A. F. REGA. 1967. Cation loading of red blood cells. J. Physiol.(Lond.). 193:459.

7. GARY-BOBO, C. M., and A. K. SOLOMON. 1968. Properties of hemoglobin solutions in redcells. J. Gen. Physiol. 52:825.

8. GLYNN, I. M. 1956. Sodium and potassium movements in human red cells. J. Physiol.(Lond.). 134:278.

9. GLYNN, I. M., and V. L. LEW. 1970. Synthesis of adenosine triphosphate at the expenseof downhill cation movements in intact human red cells. J. Physiol. (Lond.). 207:393.

10. GLYNN, I. M., V. L. LEW, and U. Lutrm. 1970. Reversal of the potassium entry mechan-ism in red cells, with and without reversal of the entire pump cycle. J. Physiol. (Lond.).207:371.

11. GLYNN, I. M., and V. Lrm. 1968. The relation between ouabain-sensitive potassiumefflux and the hypothetical dephosphorylation step in the "transport ATPase" system.J. Gen. Physiol. 51 (Suppl.):38 5s.

12. GREEN, A. L., and C. B. TAYLOR, 1964. Kinetics of sodium-potassium-stimulated adenosinetriphosphatase of rabbit kidney microsomes. Biochem. Biophys. Res. Commun. 14:118.

13. HOFFMAN, J. F. 1960. The link between metabolism and the active transport of sodium inhuman red cell ghosts. Fed. Proc. 19:127.

14. HOFFMAN, J. F. 1962 a. Cation transport and structure of the red-cell plasma membrane.Circulation. 26:1201.

372

KNIGHT AND WELT Intracellular K, a Determinant of Na-K Pump Rate

15. HOFFMAN, J. F. 1962 b. The active transport of sodium by ghosts of human red bloodcells. J. Gen. Physiol. 45:837.

16. HOFFMAN, J. F., and F. M. KREGENOW. 1966. The characterization of new energy de-pendent cation transport processes in red blood cells. Ann. N. Y. Acad. Sci. 137:566.

17. HOFFMAN, P. G. and D. C. TOSTESON. 1971. Active sodium and potassium transport inhigh potassium and low potassium sheep red cells. J. Gen. Physiol. 58:438.

18. LuBowITz, H., and R. WHITTAM. 1969. Ion movements in human red cells independentof the sodium pump. J. Physiol. (Lond.). 202:111.

19. MCCONAGHEY, P. D., and M. MAIZELS. 1962. Cation exchanges in lactose treated humanred cells. J. Physiol. (Lond.). 162:485.

20. MAIZELs, M. 1968. Effect of sodium content on sodium efflux from human red cells sus-pended in sodium-free media containing potassium, rubidium, cesium, or lithium chlo-ride. J. Physiol. (Lond.). 195:657.

21. OPIT, L. J., and J. S. CHARNocK. 1965. A molecular model of sodium pump. Nature(Lond.). 208:471.

22. POST, R. L., and P. C. JOLLY. 1957. The linkage of sodium, potassium, and ammoniumactive transport across the human erythrocyte membrane. Biochim. Biophys. Acta. 25:118

23. POST, R. L., C. R. MERRITT, C. R. KINSOLVNG, and C. D. ALBRIGHT, 1960. Membraneadenosine triphosphatase as a participant in the active transport of sodium and potas-sium in the human erythrocyte. J. Biol. Chem. 235:1796.

24. PRIESTLAND, R. N., and R. WHITTAM. 1968. The influence of external sodium ions on thesodium pump in erythrocytes. Biochem. J. 109:369.

25. ROBINSON, J. D. 1970. Interactions between monovalent cations and the sodium-potassium-dependent adenosine triphosphatase. Arch. Biochem. Biophys. 139:17.

26. SACHS, J. R. 1967. Competitive effects of some cations on active potassium transport inthe human red blood cell. J. Clin. Invest. 46:1433.

27. SACHS, J. R. 1970. Sodium movements in the human red blood cell. J. Gen. Physiol. 56:322.28. SACHS, J. R., AND L. G. WELT, 1967. The concentration dependence of active potassium

transport in the human red blood cell. J. Clin. Invest. 46:65.29. SCHATZMANN, H. J. 1953. Herzglykoside als Hemmstoffe fr den aktiven Kalium- und

Natriumtransport durch der Erythrocytenmenbran. Helv. Physiol. Pharmacol. Acta.11:346.

30. SCHATZMANN, J. J. 1965. The role of sodium and potassium in the inhibition of the sodiumpotassium-activated membrane adenosine triphosphatase. Biochim. Biophys. Acta. 94:89.

31. SKOU, J. C. 1957. The influence of some cations on an adenosine triphosphatase fromperipheral nerve. Biochim. Biophys. Acta. 23:394.

32. SKOU, J. C. 1960. Further investigations on a magnesium and sodium-activated ATPase,possibly related to the active linked transport of sodium and potassium across thenerve membrane. Biochim. Biophys. Acta. 42:6.

33. SKOU, J. C. 1965. Enzymatic basis for active transport of sodium and potassium acrosscell membranes. Physiol. Rev. 45:596.

34. SMITH, E. K. M., and L. G. WELT, 1971. Activation of human erythrocyte membraneadenosine triphosphatase by low concentrations of sodium and potassium. Am. J.Physiol. 220:119.

35. WHITTAM, R. 1962. The asymmetrical stimulation of a membrane adenosine triphosphatasein relation to active cation transport. Biochem. J. 84:110.

36. WHITTAM, R., and N. E. AGER. 1964. Vectorial aspects of the adenosine triphosphataseactivity in erythrocyte membranes. Biochem. J. 931:337.

373